Warming the fuel for the fire: Evidence for the thermal
dissociation of methane hydrate during the Paleocene-Eocene
thermal maximum
Deborah J. Thomas Department of Geological Sciences, University of North Carolina-Chapel Hill, Chapel Hill, North Carolina
27599-3315, USA
James C. Zachos Earth Sciences Department, University of California Santa Cruz, Santa Cruz, California 95064, USA
Timothy J. Bralower Department of Geological Sciences, University of North Carolina-Chapel Hill, Chapel Hill, North Carolina
27599-3315, USA
Ellen Thomas Department of Earth and Environmental Sciences, Wesleyan University, Middletown, Connecticut 06549-0139,
USA and Center for the Study of Global Change, Yale University, New Haven, Connecticut 06520-8109, USA
Steven Bohaty Earth Sciences Department, University of California Santa Cruz, Santa Cruz, California 95064, USA
ABSTRACT
Dramatic warming and upheaval of the carbon system at the end of the Paleocene Epoch
have been linked to massive dissociation of sedimentary methane hydrate. However, testing the Paleocene-Eocene thermal maximum hydrate dissociation hypothesis has been hindered by the inability of available proxy records to resolve the initial sequence of events.
The cause of the Paleocene-Eocene thermal maximum carbon isotope excursion remains
speculative, primarily due to uncertainties in the timing and duration of the PaleoceneEocene thermal maximum. We present new high-resolution stable isotope records based
on analyses of single planktonic and benthic foraminiferal shells from Ocean Drilling
Program Site 690 (Weddell Sea, Southern Ocean), demonstrating that the initial carbon
isotope excursion was geologically instantaneous and was preceded by a brief period of
gradual surface-water warming. Both of these findings support the thermal dissociation
of methane hydrate as the cause of the Paleocene-Eocene thermal maximum carbon isotope excursion. Furthermore, the data reveal that the methane-derived carbon was mixed
from the surface ocean downward, suggesting that a significant fraction of the initial
dissociated hydrate methane reached the atmosphere prior to oxidation.
Keywords: Paleocene-Eocene thermal maximum, stable isotopes, methane hydrates,
paleoceanography.
INTRODUCTION
At the close of the Paleocene Epoch ca. 55
Ma, Earth experienced one of the most dramatic global warming events in the geologic
record—the Paleocene-Eocene thermal maximum. Within a few thousand years, sea-surface temperatures warmed ;4–8 8C and deep
ocean temperatures increased by ;5 8C (Fig.
1; e.g., Kennett and Stott, 1991). The shortlived warming event (;210 k.y.; Röhl et al.,
2000) induced a host of biotic responses, including mass extinction of benthic foraminifera, rapid diversification of planktonic foraminifera and terrestrial mammals, and blooms
of dinoflagellates (e.g., Tjalsma and Lohmann,
1983; Kelly et al., 1996; Hooker, 1996; Clyde
and Gingerich, 1998; Crouch et al., 2001).
The Paleocene-Eocene thermal maximum
corresponds to an abrupt negative 3‰–4‰
carbon isotope excursion in marine and terrestrial sedimentary sections (Fig. 1; e.g.,
Kennett and Stott, 1991; Koch et al., 1992), a
feature with major diagnostic significance.
This excursion implies a massive and rapid
addition of isotopically light carbon to the
oceans and atmosphere. Possible explanations
Figure 1. Carbon and oxygen isotope records from Ocean Drilling Program Site 690
(data from Kennett and Stott, 1991) showing
Paleocene-Eocene thermal maximum stable
isotope excursions.
for this input have focused on enhanced mantle CO2 outgassing during North Atlantic Igneous Province emplacement (Eldholm and
Thomas, 1993), dissociation of massive quantities of methane hydrates along continental
slopes (Dickens et al., 1995, 1997), and impact of a carbonaceous bolide (Deming, 1999;
Kent et al., 2001). The apparent rate and magnitude of the carbon isotope excursion are
consistent with both seafloor methane and a
carbonaceous impactor as the carbon sources.
However, the available data cannot distinguish
between these two sources of light carbon, or
whether the carbon (methane or otherwise)
was added first to the oceans or atmosphere.
The fundamental problem is that temporal relationships between warming and carbon input
in different environments cannot be resolved
with published stable isotope records.
We present new stable isotope data from
Ocean Drilling Program Site 690 (Weddell
Sea, Southern Ocean) that indicate that the onset of the carbon isotope excursion was geologically instantaneous, consistent with the
hydrate dissociation hypothesis or impactrelated carbon release. However, the data also
demonstrate that the carbon isotope excursion
at Site 690 was preceded by a brief period of
gradual surface-water warming. Thus the most
coherent explanation for the entire data set is
that gradual sea-surface warming led to the
thermal dissociation of sedimentary methane
hydrate. In addition, our data suggest that a
substantial amount of the methane released
was oxidized in the atmosphere.
METHODS
We generated stable isotope records based
on analyses of individual foraminiferal shells,
a strategy used in two previous PaleoceneEocene thermal maximum investigations
(Stott, 1992; Kelly et al., 1996). Site 690 (Fig.
2) was selected because it is the most expanded and complete deep-sea Paleocene-Eocene
thermal maximum section recovered to date
(Röhl et al., 2000). A continuous U-channel
q 2002 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or editing@geosociety.org.
Geology; December 2002; v. 30; no. 12; p. 1067–1070; 3 figures.
1067
ern Ocean thermal and carbon isotope depth
gradients during the onset and initial recovery
of the Paleocene-Eocene thermal maximum.
Kennett and Stott (1991) first recognized the
stable isotope excursions at Site 690; however,
their multispecimen analyses lacked the resolution to unravel the timing relationships between warming and carbon input. Multispecimen records also average out the
environmental variability that the singlespecimen strategy exploits. As a result, we are
now able to resolve the onset and evolution of
the thermal and carbon anomalies. Our data
support a warming-induced carbon input such
as thermal dissociation of methane hydrates.
Figure 2. Paleogeographic reconstruction of late Paleocene showing location of Site 690
(from Ocean Drilling Stratigraphic Network).
across the onset of the Paleocene-Eocene thermal maximum was cut into 1 cm samples;
above this we sampled the core every 5 cm.
We conducted stable isotope analyses on 483
well-preserved specimens (Thomas et al.,
1999) of the surface-dwelling, photosymbiontbearing genus Acarinina and the thermoclinedwelling genus Subbotina (D’Hondt and Zachos, 1998) over a 121 cm interval, and on 7
individual and 5 multispecimen benthic foraminiferal samples over a 7 cm interval (ending
4 cm below the benthic foraminiferal extinction). Measurements of individual benthic foraminifera N. truempyi were limited because
specimens large enough (.5 mg) for analysis
were extremely rare, especially close to the
onset of the benthic foraminiferal extinction.
All stable isotope analyses were conducted at
the University of California Santa Cruz with
a precision better than 0.1‰ for d13C and
d18O.
RESULTS
Several key features emerge from the
single-specimen stable isotope data (Fig. 3).
Specimens of surface- and thermoclinedwelling planktonic foraminifera record only
preexcursion or excursion d13C values across
the onset of the event. The onset of excursion
d13C values within specimens of both groups
occurs within a span of 1 cm, and coincides
with individuals that still record pre–carbon
isotope excursion values. In contrast to the bimodal planktonic carbon isotope distribution,
both the surface- and thermocline-dwelling foraminiferal d18O records contain intermediate
values, marking a more gradual transition
from pre–Paleocene-Eocene thermal maximum to Paleocene-Eocene thermal maximum
d18O values. Transitional d18O values are expressed by the same individual specimens that
contain pre–Paleocene-Eocene thermal maximum d13C values.
1068
Shifts in stable isotope values of different
groups of foraminifera also show a stratigraphic sequence of events (Fig. 3). From bottom to top, these are (1) d18O values of
surface-dwelling planktonics decrease by
;1.5‰ between 170.81 m below seafloor
(mbsf; level 1) and 170.78 mbsf (level 2). (2)
Level 2 also corresponds to the initial decrease
in d18O values of thermocline-dwelling planktonics and the ;4‰ decrease in d13C values
of surface-dwelling specimens. (3) At 170.70
mbsf (level 3), d13C values of thermocline dwellers decrease by ;2.5‰. (4) Peak PaleoceneEocene thermal maximum d18O values of
thermocline dwellers (;22‰) occur at level
4 (170.67 mbsf). Level 4 also marks a slight
decrease in d13C (;1.5‰) and d18O values
(;1‰) of two benthic individuals. New multispecimen and single-specimen benthic analyses above these two individuals indicate reworking based on pre–Paleocene-Eocene
thermal maximum d13C and d18O values.
Above a brief convergence at the onset of
the Paleocene-Eocene thermal maximum
(from 170.70 to 170.42 mbsf), the trends in
planktonic stable isotope values begin to diverge. Oxygen isotope values of thermocline
dwellers remain at ;21.5‰ up to 170.42
mbsf, then begin to gradually increase. However, d13C values of thermocline dwellers remain low throughout the interval studied (to
169.84 mbsf). Oxygen isotope values of surface dwellers are ;21.5‰ from 170.78 to
170.42 mbsf, decrease from ;21.5 to 22.5‰
at 170.37 mbsf, and remain at ;22.5‰ for
the remainder of the record. Carbon isotope
values of surface dwellers begin to increase at
170.37 mbsf.
DISCUSSION
Our combined carbon and oxygen singlespecimen isotopic data (Fig. 3) provide a detailed chronology of rapid changes in South-
Evidence for the Blast of Gas
The bimodal distribution of singlespecimen carbon isotope values at the onset of
the carbon isotope excursion suggests that the
initial input of isotopically light carbon occurred in a geologic instant. We can discount
the argument that the sharp base of the carbon
isotope excursion resulted from dissolution or
a hiatus (Dickens, 1998), because the very
planktonic foraminifera that record only preor peak excursion d13C values record a gradual decrease in the corresponding d18O values.
The geologically instantaneous onset of the
carbon isotope excursion precludes the North
Atlantic Igneous Province as the source of
light carbon, because the d13C value of mantle
CO2 is only ;26‰. The stratigraphic and hydrographic sequence of the first excursion
d13C values from surface to thermocline to
benthic dwellers implies that the carbon isotope anomaly gradually propagated downward
from the sea surface to the bottom waters.
Triggering the Blast
The stratigraphic progression of singlespecimen stable isotope changes, with the decrease in surface-water d18O values preceding
the decrease in d13C values, enables us to rule
out several possible explanations of
Paleocene-Eocene thermal maximum carbon
input. The onset of the carbon isotope excursion would have preceded the decrease in
d18O values if the Paleocene-Eocene thermal
maximum had resulted from erosion or impactinduced hydrate dissociation (e.g., Katz et al.,
2001; Kent et al., 2001). Explosive volcanism
(e.g., Bralower et al., 1997) would have resulted in an increase or no change in d18O
values at a high-latitude site. Thus, the most
plausible mechanism to consider is the thermal dissociation of methane hydrates.
The occurrence of specimens of surfacedwelling foraminifera that record transitional
d18O values and pre–carbon isotope excursion
d13C values (level 1, Fig. 3) suggests an ;2
8C warming of surface waters prior to the on-
GEOLOGY, December 2002
bility. With the onset of the carbon isotope
excursion (level 2, Fig. 3) surface mixed-layer
temperatures eventually warmed an additional
4 8C. Transitional d18O values recorded by
thermocline dwellers occur above the d18O decrease in surface dwellers (level 2, Fig. 3), indicating a distinct delay in the warming of the
thermocline waters (several hundred meters
depth). Moreover, the high degree of scatter in
the transitional thermocline-dweller d18O data
is consistent with the notion that thermocline
structure changed, prompting shifts in foraminiferal depth habitats as the intermediate
and deep ocean began to warm.
Figure 3. Single-specimen carbon and oxygen isotope data. Previously published benthic
foraminiferal data are from Kennett and Stott (1991) and Thomas and Shackleton (1996).
Gray arrow denotes onset of benthic foraminiferal extinction. Gray dashed lines indicate
stratigraphic sequence of events at onset of Paleocene-Eocene thermal maximum. Highresolution samples derived from archive-half U-channel of Hole 690B, section 19-3, extend
from 170.40 to 171.11 m below seafloor (mbsf). Above and below U-channel (169.84–171.73
mbsf), samples were taken from archive half at 2–5 cm intervals. Individuals from genera
Acarinina and Subbotina, as well as N. truempyi were handpicked from .250 mm size fraction. PDB is Peedee belemnite.
set of the excursion. This finding conflicts
with that of Bains et al. (1999), who found no
evidence for pre–Paleocene-Eocene thermal
maximum sea-surface warming on the basis of
bulk sedimentary oxygen isotope analyses.
We argue that the discrepancy results from the
different analytical techniques, because chang-
GEOLOGY, December 2002
es in bulk stable isotopes at Site 690 during
the Paleocene-Eocene thermal maximum interval reflect changes in the assemblage of calcareous nannofossils (Bralower, 2002). The
cause of this warming is not known, although
long-term outgassing of CO2 from the North
Atlantic Igneous Province is a distinct possi-
Pathway of Hydrate Carbon
The top-down progression of the onset of
the carbon isotope excursion suggests that a
significant proportion of the methane from
dissociated hydrates was rapidly transferred to
the atmosphere and surface ocean. In order for
calcifying organisms to record a methanederived d13C anomaly, the isotopically light
methane must first be oxidized into CO2 and
2
incorporated into the HCO3 pool from which
calcification occurs. Because the pattern of
carbon isotope excursion propagation proceeded downward from surface waters, oxidation of methane must have taken place within the atmosphere–surface ocean. Had the
initial release of methane been more gradual
(enabling oxidation within the deep ocean),
Site 690 planktonic foraminifera would have
recorded transitional d13C values at the onset
of the event, and benthic individuals would
have recorded the excursion prior to the
planktonics.
At Site 690, progression of the initial d18O
decrease also occurs from surface waters
downward. Heat may have been mixed vertically down to the thermocline, while warmer
intermediate and deep waters may have been
advected into the study region after sufficient
warming in their respective source areas. The
onset of peak thermocline warming (lowest
d18O values) coincides with the onset of the
carbon isotope excursion in surface waters.
This implies that hydrate dissociation did not
occur at Site 690, otherwise thermocline and
deeper water warming would have preceded
the onset of the carbon isotope excursion.
The more gradual transition to excursion
d13C values in the benthic foraminiferal record
can be understood by considering the dynamics of mixing and the vast size of the deepsea inorganic C reservoir. As indicated by the
bimodal planktonic d13C data, the mean d13C
of the relatively small and well-mixed atmosphere and surface-ocean carbon reservoirs
should have responded immediately to the initial release of methane. In contrast, because of
the large mass of carbon and the slower mix-
1069
ing time of the deep ocean (;1500 yr), it
should take several thousands of years for the
full magnitude of the carbon isotope excursion
to be manifested in the deep sea. Thus, the
delay in the thermocline and benthic d13C decrease reflects the additional time required to
mix the carbon anomaly throughout the deepsea carbon reservoir. The lack of benthic individuals available for analysis (during the decline and extinction of benthic foraminifera)
may also bias the timing of environmental
change in the deep sea.
We note that the top-down progression in
carbon input observed at the PaleoceneEocene thermal maximum is strikingly similar
to the recent changes in the atmospheric and
surface ocean carbon reservoirs in response to
release of anthropogenic CO2. In the past
century, the d13C of atmospheric and surfaceocean CO2 has decreased by ;1.2‰ (e.g.,
Joos and Bruno, 1998), while the deep ocean
has yet to show detectable change.
Paleocene-Eocene Thermal Maximum
Recovery
Increasing surface-dwelling planktonic
d13C values suggest that initial recovery of
surface-water dissolved inorganic carbon
(DIC) began even while thermocline and benthic foraminifera still recorded carbon isotope
excursion values. The smaller surface-ocean
carbon reservoir should have recovered from
the isotopic perturbation more rapidly than the
larger deeper oceanic reservoirs. However,
that the surface dwellers do not record any
subsequent d13C decrease after the initial onset of the carbon isotope excursion implies
that all of the hydrate dissociation may have
occurred in one major event. Additional hydrate dissociation should have resulted in further decrease of surface-water DIC d13C values. Alternatively, subsequent gradual hydrate
may have occurred, but the methane-derived
carbon remained within the deep-ocean
reservoir.
The thermal structure of the Southern
Ocean water column changed markedly during
the several tens of thousands years following
the onset of the Paleocene-Eocene thermal
maximum. After an initial convergence of surface and thermocline d18O values at the onset
of the warming (as noted by Kennett and
Stott, 1991), thermal stratification resumed.
Coincident with the initial recovery in
surface-water d13C values, surface-water d18O
values decreased by another ;1‰, possibly
as a consequence of enhanced greenhouse
warming. These trends suggest a reestablish-
1070
ment of local
stratification.
thermal
and
chemical
PROPOSED CAUSAL SCENARIO FOR
THE PALEOCENE-EOCENE
THERMAL MAXIMUM
We propose the following scenario to explain the stratigraphic sequence of events in
the new stable isotope data. Gradual warming
occurred first in surface waters, then in waters
at thermocline and intermediate depths. Subduction or downwelling of warmer intermediate waters in the region of water-mass formation led to thermal dissociation of methane
hydrates at a location with a significant sedimentary hydrate content. Methane gas from
the dissociated hydrates reached the atmosphere prior to widespread oxidation. The
thermocline–intermediate water thermal
anomaly likely reached the Southern Ocean
via a combination of advection from a distal
source region and vertical diffusion. The propagation of the carbon anomaly proceeded rapidly throughout the atmosphere and surface
ocean, then more slowly downward into the
thermocline and then bottom waters.
ACKNOWLEDGMENTS
We thank Paula Weiss and the East Coast Repository
staff for helping procure the 690 U-channel, and the
Ocean Drilling Program. We also thank Jerry Dickens
and Steve D’Hondt for thoughtful reviews that strengthened this manuscript, as well as an anonymous reader.
National Science Foundation grant EAR-98-14604 to
Bralower and Zachos funded this project.
REFERENCES CITED
Bains, S., Corfield, R.M., and Norris, R.D., 1999,
Mechanisms of climate warming at the end of the
Paleocene: Science, v. 285, p. 724–727.
Bralower, T.J., 2002, Evidence for surface water oligotrophy during the late Paleocene thermal maximum: Nannofossil assemblage data from Ocean
Drilling Program Site 690, Maud Rise, Weddell
Sea: Paleoceanography (in press).
Bralower, T.J., Thomas, D.J., Zachos, J.C., Hirschmann,
M.M., Rohl, U., Sigurdsson, H., Thomas, E., and
Whitney, D.L., 1997, High-resolution records of
late Paleocene thermal maximum and circumCaribbean volcanism: Is there a causal link?: Geology, v. 25, p. 963–966.
Clyde, W.C., and Gingerich, P.D., 1998, Mammalian
community response to the latest Paleocene thermal maximum: An isotaphonomic study in the
northern Bighorn Basin, Wyoming: Geology,
v. 26, p. 1011–1014.
Crouch, E., Heilmann-Clausen, C., Brinkhuis, H., Morgans, H.E.G., Rogers, K.M., Egger, H., and
Schmitz, B., 2001, Global dinoflagellate event associated with the late Paleocene thermal maximum: Geology, v. 29, p. 289–376.
Deming, D., 1999, On the possible influence of extraterrestrial volatiles on Earth’s climate and the origin of the oceans: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 146, p. 33–51.
D’Hondt, S., and Zachos, J.C., 1998, Cretaceous Foraminifera and the evolutionary history of planktic photosymbiosis: Paleobiology, v. 24,
p. 512–523.
Dickens, G.R., 1998, High-resolution records of the late
Paleocene thermal maximum and circum-Caribbean
volcanism: Is there a causal link?: Comment: Geology, v. 26, p. 670.
Dickens, G.R., O’Neil, J.R., Rea, D.C., and Owen,
R.M., 1995, Dissociation of oceanic methane hydrate as a cause of the carbon isotope excursion
at the end of the Paleocene: Paleoceanography,
v. 10, p. 965–971.
Dickens, G.R., Castillo, M.M., and Walker, J.C.G.,
1997, A blast of gas in the latest Paleocene: Simulating first-order effects of massive dissociation
of oceanic methane hydrate: Geology, v. 25,
p. 258–262.
Eldholm, O., and Thomas, E., 1993, Environmental impact of volcanic margin formation: Earth and
Planetary Science Letters, v. 117, p. 319–329.
Hooker, J.J., 1996, Mammalian biostratigraphy across
the Paleocene-Eocene boundary in the Paris, London and Belgian basins, in Knox, R.W.O., et al.,
eds., Correlation of the early Paleogene in Northwest Europe: Geological Society [London] Special Publication 101, p. 205–218.
Joos, F., and Bruno, M., 1998, Long-term variability of
the terrestrial and oceanic carbon sinks and the
budgets of the carbon isotopes 13C and 14C: Global Biogeochemical Cycles, v. 12, p. 277–295.
Katz, M.E., Cramer, B.S., Mountain, G.S., Katz, S., and
Miller, K.G., 2001, Uncorking the bottle: What
triggered the Paleocene/Eocene thermal maximum methane release?: Paleoceanography, v. 16,
p. 549–562.
Kelly, D.C., Bralower, T.J., Zachos, J.C., Premoli Silva,
I., and Thomas, E., 1996, Rapid diversification of
planktonic foraminifera in the tropical Pacific
(ODP Site 865) during the late Paleocene thermal
maximum: Geology, v. 24, p. 423–426.
Kennett, J.P., and Stott, L.D., 1991, Abrupt deep-sea
warming, palaeoceanographic changes and benthic extinctions at the end of the Palaeocene: Nature, v. 353, p. 225–229.
Kent, D.V., Cramer, B.S., and Lanci, L., 2001, Evidence
of an impact trigger for the Paleocene/Eocene
thermal maximum and carbon isotope excursion:
Eos (Transactions, American Geophysical
Union), v. 82, Fall Meeting Supplement, Abstract
PP32A-0509.
Koch, P.L., Zachos, J.C., and Gingerich, P.D., 1992,
Correlation between isotopic records in marine
and continental reservoirs near the Paleocene/Eocene boundary: Nature, v. 358, p. 319–322.
Röhl, U., Bralower, T.J., Norris, R.N., and Wefer, G.,
2000, A new chronology for the late Paleocene
thermal maximum and its environmental implications: Geology, v. 28, p. 927–930.
Stott, L.D., 1992, Higher temperatures and lower oceanic pCO2: A climate enigma at the end of the
Paleocene Epoch: Paleoceanography, v. 7,
p. 395–404.
Thomas, E., and Shackleton, N.J., 1996, The latest Paleocene benthic foraminiferal extinction and stable isotope anomalies, in Knox, R.W.O., et al.,
eds., Correlation of the early Paleogene in Northwest Europe: Geological Society [London] Special Publication 101, p. 401–441.
Thomas, D.J., Bralower, T.J., and Zachos, J.C., 1999,
New evidence for subtropical warming during the
late Paleocene thermal maximum: Stable isotopes
from Deep Sea Drilling Project Site 527, Walvis
Ridge: Paleoceanography, v. 14, p. 561–570.
Tjalsma, R.C., and Lohmann, G.P., 1983, PaleoceneEocene bathyal and abyssal benthic foraminifera
from the Atlantic Ocean: Micropaleontology,
Special Publication 4, 90 p.
Manuscript received May 7, 2002
Revised manuscript received August 5, 2002
Manuscript accepted August 13, 2002
Printed in USA
GEOLOGY, December 2002